Gradual Study Advising with Course Knowledge Graphs

Knowledge graphs (KGs) have been actively studied for pedagogical purposes. To depict the rich but latent relations among different concepts in the course textbook, increasing efforts have been proposed to construct course KGs for university students. However, the application of course KGs for real study scenarios and career development remains unexplored and nontrivial. First, it is hard to enable personalized viewing and advising. Within the intricate university curricula, instructors aim to assist students in developing a personalized course selection pathway, which cannot be fulfilled by isolated course KGs. Second, locating concepts that are important to individuals poses challenges to students. Real-world course KGs may contain hundreds of concepts connected by hierarchical relations, e.g., contain_subtopic, making it challenging to capture the key points. To tackle these challenges, in this paper, we present GSA, a novel gradual study advising system based on course knowledge graphs, to facilitate both intra-course study and inter-course development for students significantly. Specifically, (i) we establish an interactive web system for both instructors to construct and manipulate course KGs, and students to view and interact. (ii) Concept-level advising is designed to visualize the centrality of a course KG based on various metrics. We also propose a tailored algorithm to suggest the learning path based on what concepts students have learned. (iii) Course-level advising is instantiated with a course network. This indicates the prerequisite relation among different levels of courses, corresponding to the annually increasing curricular design and forming different major streams. Extensive illustrations show the effectiveness of our system.